Modelling survival time of cervical cancer patients in Moi Teaching and Referral Hospital, Kenya

Abstract

Cervical cancer is one of the most prevalent cancers in women. Globally, it is ranked behind breast cancer, lung cancer, and colorectal cancer as the fourth most common cancer killer of females. Cervical cancer has risen 3,211 percent in the last five years, making it the top cancer-related death among women in Kenya. An increase in cervical cancer in Kenya has resulted in an economic burden for patients and families. There is an increase in healthcare spending and productivity losses due to morbidity and mortality at a productive age. The purpose of the study was to model the survival time of cervical cancer patients at the Moi Teaching and Referral Hospital (MTRH) to gain insights into causes and prognostic factors of survival after diagnosis. The specific objectives were to (1) estimate the survival time among cervical cancer patients, (2) determine the predictors of survival after a cervical cancer diagnosis, and (3) evaluate the relationship between a patient's survival and covariates of cervical cancer. The study employed a retrospective cohort design. A population of 175 cervical cancer patients enrolled between 1st January to 31st December 2014 were studied for a follow-up period of five years. Objective one was achieved using the Kaplan-Meier estimator to estimate the median survival time. The Cox proportional hazard regression model was used to achieve objectives two and three. Data analysis was done using R statistical software. The study findings revealed that 144(82.3%) of cervical cancer patients who were considered in the study died within 5 years. Furthermore, 98.6% of patients who had a family history of cervical cancer died within the study period. Only 33.3% (10) of patients diagnosed with stage one of cervical cancer died within the study period. Furthermore, 81.6% and 94.8% of patients diagnosed with stage two and three cervical cancer, respectively, died within the study period. Finally, all patients diagnosed with stage four died within five years. Among the patients whose treatment plan was radiation, 56.4% of them died within the study period across all stages of the disease. The age of the patients [hazard ratio (HR) =0.369; p-value (p) =<0.001], employment status [HR =.328, p=0.042], family history of cervical cancer [HR=0.444, p=0.048], smoking cigarettes [HR=0.807, p=0.002], comorbidities [HR =0.825, p=0.001], cancer grade [HR =.472, p=0.001], staging of cervical cancer [HR =0.265, p=0.015] and treatment plan [HR =-0.124, p=0.043] were significant predictors for survival time of cervical cancer patients. The overall median survival time was two years. The study concluded that age, marital status, employment status, family history, smoking status, comorbidity, cancer grade, staging of the disease and treatment plan were significant predictors of survival after a diagnosis of cervical cancer. However, insurance cover, attaining menopause age, use of contraceptives, types of contraceptives, HIV status and HPV infections were not significant predictors of survival after a diagnosis of cervical cancer. There was an increased chance of survival if cervical cancer patients were diagnosed in stage one. The study recommends that healthcare systems in MTRH and Kenya should consider investing in highly targeted social support services and interventions that may help reduce the significant survival differences due to predictors of survival after a diagnosis.